Abstract 34

Background: Long-term daily consumption of fruits and vegetables may prevent or reduce a number of chronic diseases, including certain cancers, obesity, and diabetes. However, consumption of fruits and vegetables among children and adolescents consistently falls below the 5-a-Day recommendations.

Objective: To identify modifiable predictors of fruit and vegetable consumption among a representative sample of American adolescents.

Methods: We conducted a secondary analysis using data collected through the National Longitudinal Study of Adolescent Health (AddHealth). This representative population-based sample of high school students completed an in-home survey, providing detailed information about health-related behaviors, including a one day recall of fruit and vegetable consumption. We used the AddHealth data to evaluate previously reported predictors of fruit and vegetable consumption, including race/ethnicity, gender, level of physical activity, time spent watching television and engaging in other sedentary activities, weekly fast food consumption, and intentions to lose or gain weight. Bivariate tests of association and multivariate regression were used to evaluate the relationship between these predictors and reported fruit and vegetable consumption.

Results: A total of 4832 high school students (52% female) completed the survey. Sixty-six percent of the males and 67% of the females consumed fewer than 5 servings of fruits and vegetables per day. Among the minority ethnic groups, African-American adolescents consumed the fewest servings of fruits and vegetables, with 71% consuming fewer than 5 servings per day. Prevalence of fast-food consumption was high among all teens, with 57% consuming 2 or more fast food meals per week. Physical activity, family meal patterns and peer group influence on fruit and vegetable consumption were also noted.

Conclusion: Patterns of poor fruit and vegetable consumption among adolescents transcend ethnic, and gender barriers. Identification of the modifiable predictor variables can facilitate the design of effective interventions targeting this at-risk population.